Method and device for determining promoters, medium, electronic equipment and program product

Abstract

A method and device for determining a promoter, a medium, an electronic device and a program product, relating to the technical field of Internet, the method comprising: obtaining a first vector corresponding to a target delivery party and a second vector corresponding to at least one candidate promoter from a pre-trained graph model, the graph model is generated based on historical interaction records between every two of at least one delivery party, at least one candidate promotion party and at least one user, and is used for describing the delivery parties, the candidate promotion parties and the users through vectors, and the at least one delivery party comprises a target delivery party; determining the similarity between each second vector and the first vector corresponding to the target delivery party; and determining a target promotion party of the target delivery party from the plurality of candidate promotion parties based on the similarity. The graph model fully mines the similarity among the delivery party, the candidate promoters and the user, so that the best promoters can be accurately selected for the delivery party based on the graph model, and the effectiveness of information promotion is improved.

Description

Technical Field

[0001] The present disclosure relates to the field of Internet technologies, and in particular, to a method, an apparatus, a medium, an electronic device, and a program product for determining a promoter. Background Art

[0002] With the development of the Internet, the Internet has become the main channel for advertisers to promote information, such as the promotion of advertisements by advertisers.

[0003] In the related art, advertisers generally use the accounts provided by promoters with certain influence to promote information. With the development of the Internet, the number of promoters with certain influence is increasing. Therefore, how to select the best promoter for advertisers from a large number of promoters is a technical problem that needs to be solved urgently at present. Summary of the Invention

[0004] This Summary of the Invention section is provided to introduce concepts in a brief form, which will be described in detail in the following Detailed Implementation section. This Summary of the Invention section is not intended to identify key features or essential features of the claimed technical solution, nor is it intended to be used to limit the scope of the claimed technical solution.

[0005] In a first aspect, the present disclosure provides a method for determining a promoter, including: Obtaining a first vector corresponding to a target advertiser and second vectors corresponding to at least one candidate promoter from a pre-trained graph model, where the graph model is generated based on historical interaction records between any two of at least one advertiser, at least one candidate promoter, and at least one user, and is used to describe each advertiser, each candidate promoter, and each user through vectors, and the at least one advertiser includes the target advertiser; Determining the similarity between each of the second vectors and the first vector corresponding to the target advertiser; Based on the similarity, determining the target promoter of the target advertiser from the multiple candidate promoters.

[0006] In a second aspect, the present disclosure provides an apparatus for determining a promoter, including: An obtaining module, configured to obtain a first vector corresponding to a target advertiser and second vectors corresponding to at least one candidate promoter from a pre-trained graph model, where the graph model is generated based on historical interaction records between any two of at least one advertiser, at least one candidate promoter, and at least one user, and is used to describe each advertiser, each candidate promoter, and each user through vectors, and the at least one advertiser includes the target advertiser; A first determination module, configured to determine the similarity between each of the second vectors and the first vector corresponding to the target advertiser. A second determination module, configured to determine the target advertiser of the target advertiser from the multiple candidate promoters based on the similarity.

[0007] In a third aspect, the present disclosure provides a computer-readable medium, on which a computer program is stored, and when the computer program is executed by a processing device, the steps of the method described in the first aspect are implemented.

[0008] In a fourth aspect, the present disclosure provides an electronic device, including: A storage device, on which a computer program is stored; A processing device, configured to execute the computer program in the storage device to implement the steps of the method described in the first aspect.

[0009] In a fifth aspect, the present disclosure provides a computer program product, including a computer program, and when the computer program is executed by a processor, the steps of the method described in the first aspect are implemented.

[0010] Through the above technical solutions, the first vector corresponding to the target advertiser and the second vectors corresponding to at least one candidate promoter are obtained from a pre-trained graph model, and the similarity between each second vector and the first vector corresponding to the target advertiser is determined; based on the similarity, the target promoter of the target advertiser is determined from multiple candidate promoters. Since the graph model can fully exploit the similarities between advertisers, candidate promoters, and users, based on the graph model, the best promoter can be accurately selected for the advertiser, thereby improving the effectiveness of information promotion.

[0011] Other features and advantages of the present disclosure will be described in detail in the subsequent specific implementation section. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In combination with the accompanying drawings and with reference to the following specific implementation, the above and other features, advantages, and aspects of the various embodiments of the present disclosure will become more apparent. Throughout the drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic, and the original components and elements are not necessarily drawn to scale. In the drawings: Figure 1 is a schematic diagram of an application scenario shown according to an embodiment of the present disclosure.

[0013] Figure 2 is a flowchart of a method for determining a promoter shown according to an embodiment of the present disclosure.

[0014] Figure 3 is a schematic diagram of a graph network shown according to an embodiment of the present disclosure.

[0015] Figure 4 It is a block diagram of a device for determining a promoter shown according to an embodiment of the present disclosure.

[0016] Figure 5 It is a schematic structural diagram of an electronic device shown according to an embodiment of the present disclosure. Detailed implementation manners

[0017] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although some embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. On the contrary, these embodiments are provided to more thoroughly and completely understand the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not used to limit the protection scope of the present disclosure.

[0018] It should be understood that the steps recited in the method embodiments of the present disclosure can be executed in different orders and / or in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present disclosure is not limited in this regard.

[0019] As used herein, the term "including" and its variations are open-ended, that is, "including but not limited to". The term "based on" means "at least partially based on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". The relevant definitions of other terms will be given in the following description.

[0020] It should be noted that the concepts such as "first" and "second" mentioned in the present disclosure are only used to distinguish different devices, modules or units, and are not used to limit the order of functions performed by these devices, modules or units or their interdependent relationships.

[0021] It should be noted that the modifications of "one" and "multiple" mentioned in the present disclosure are illustrative rather than restrictive. Those skilled in the art should understand that unless otherwise clearly specified in the context, it should be understood as "one or more".

[0022] The names of the messages or information exchanged between multiple devices in the embodiments of the present disclosure are only for illustrative purposes and are not used to limit the scope of these messages or information.

[0023] It can be understood that before using the technical solutions disclosed in the embodiments of the present disclosure, the types, usage scopes, usage scenarios, etc. of the personal information involved in the present disclosure should be informed to users and the authorization of users should be obtained in an appropriate manner in accordance with relevant laws and regulations.

[0024] For example, when receiving an active request from a user, a prompt message is sent to the user to clearly prompt the user that the operation requested by the user will require obtaining and using the user's personal information. Thus, the user can autonomously choose whether to provide personal information to software or hardware such as an electronic device, an application, a server, or a storage medium that performs the operations of the present disclosure's technical solution based on the prompt message.

[0025] As an optional but non-limiting implementation manner, the manner of sending a prompt message to the user in response to receiving an active request from the user may be, for example, in the form of a pop-up window, and the prompt message may be presented in text in the pop-up window. In addition, the pop-up window may also carry a selection control for the user to choose "agree" or "disagree" to provide personal information to the electronic device.

[0026] It can be understood that the above notification and obtaining user authorization process is only illustrative and does not limit the implementation manner of the present disclosure. Other manners that meet relevant laws and regulations can also be applied to the implementation manner of the present disclosure.

[0027] At the same time, it can be understood that the data involved in the present technical solution (including but not limited to the data itself, the acquisition or use of the data) should comply with the requirements of corresponding laws, regulations and related provisions.

[0028] In the related art, a promoter is recommended to the advertiser through industry matching. For example, the content published by the promoter and the content that the advertiser wants to promote are matched, and a promoter who has published content similar to the content that the advertiser wants to promote is recommended to the advertiser. However, this method will not expand the advertiser to promoters in other industries, and even if promoters in the same industry are recommended to the advertiser, it is not easy to distinguish the differences between the promoters.

[0029] In view of this, embodiments of the present disclosure propose a method, an apparatus, a medium, an electronic device, and a program product for determining a promoter. The present disclosure will be further explained and illustrated below with reference to the accompanying drawings.

[0030] Figure 1 is a schematic diagram of an application scenario shown according to an embodiment of the present disclosure. In Figure 1 where N, M, and K are all positive integers greater than or equal to 3. In Figure 1Among them, the advertisers 1, 2, …, N can be different merchants, and push information (such as the products of the merchants) to the users corresponding to the user devices 1, 2, …, K through the candidate promoters 1, 2, …, M. For example, an electronic device can be used to determine, according to a pre-trained graph model, which candidate promoters the target advertiser (such as advertiser 1) uses to push information to user devices. The electronic device can be, for example, a hardware device such as a computer, a tablet computer, or a smart phone, or an application running on the hardware device, or a merchant client or server of an e-commerce platform. There is no limitation in this implementation scenario.

[0031] In this embodiment, the attention degree of the candidate promoter is greater than the preset attention degree. The candidate promoters usually focus on content creation, live streaming, or social media interaction for information promotion. Taking products as an example, the candidate promoters can create content related to the products and publish the content to the Internet that can reach the users corresponding to the user devices. As an example, the created content can be videos, articles, etc.

[0032] In this embodiment, the attention degree of the candidate promoter can be determined according to the number of users of the account provided by the attention promoter for promoting information. For example, if the number of users following the account is greater than or equal to the preset number, it indicates that the attention degree of the account is greater than the preset attention degree; if the number of users following the account is less than the preset number, it also indicates that the attention degree of the account is greater than the preset attention degree. Further, the preset number can be set according to the actual situation, and there is no limitation in this embodiment.

[0033] In Figure 1 the shown implementation scenario, the electronic device can be implemented based on a pre-trained graph model. For the training of the graph model, reference can be made to the following related embodiments, which will not be elaborated in this embodiment.

[0034] In the following embodiments, for the convenience of description and understanding, the candidate promoters can be referred to as influencers, and the advertisers can be referred to as merchants.

[0035] Figure 2 is a flowchart of a method for determining a promoter according to an embodiment of the present disclosure. The method for determining a promoter can be applied to an electronic device, and the method for determining a promoter can be executed by a device for determining a promoter, where the device for determining a promoter can be implemented by software and / or hardware, and the software and / or hardware can be configured in, for example, the above-mentioned electronic device. Referring to Figure 2 this, the method for determining a promoter can include step 210, step 220, and step 230.

[0036] In step 210, obtain the first vector corresponding to the target advertiser and the second vectors corresponding to at least one candidate promoter respectively from a pre-trained graph model, where the graph model is generated based on the historical interaction records between any two of at least one advertiser, at least one candidate promoter, and at least one user, and is used to describe each advertiser, each candidate promoter, and each user through vectors, and at least one advertiser includes the target advertiser; In step 220, determine the similarity between each second vector and the first vector corresponding to the target advertiser; In step 230, based on the similarity, determine the target promoter of the target advertiser from multiple candidate promoters.

[0037] Through the above technical solution, obtain the first vector corresponding to the target advertiser and the second vectors corresponding to at least one candidate promoter respectively from a pre-trained graph model, determine the similarity between each second vector and the first vector corresponding to the target advertiser; based on the similarity, determine the target promoter of the target advertiser from multiple candidate promoters. Since the graph model can fully exploit the similarities between advertisers, candidate promoters, and users, therefore, based on the graph model, the best promoter can be accurately selected for the advertiser, thereby improving the effectiveness of information promotion.

[0038] In some embodiments, the historical interaction records may include the interaction records between each candidate promoter and each user, the cooperation records between each candidate promoter and each advertiser, and the conversion records between each advertiser and each user.

[0039] For example, if there is a conversion behavior (such as a purchase behavior) of a user for the goods of an advertiser, a corresponding conversion record can be generated; for another example, if there is a cooperation behavior between a candidate promoter and an advertiser, a corresponding cooperation record can be generated; for another example, if a user views the information released by a candidate promoter, a corresponding viewing record can be generated.

[0040] In some embodiments, the similarity may be a cosine similarity, and the cosine value of the angle between two vectors in the vector space can be used as a measure of the difference between the two vectors, that is, the magnitude of the similarity. For example, the cosine value (similarity) between the first vector Vm corresponding to the target advertiser m and the second vector Vj corresponding to the candidate promoter j can be determined by the following formula (1): (1) In the above formula (1), is the cosine value between the first vector Vm and the second vector Vj, is the norm of the first vector Vm, is the norm of the second vector Vj.

[0041] In some embodiments, step 230 above may be implemented in the following manner: sort all similarities from high to low to obtain a sorting result; use the candidate promoters corresponding to the similarities in the top preset positions in the sorting result as the target promoters.

[0042] Among them, the preset position can be set according to the actual situation, and this embodiment does not limit it.

[0043] In some embodiments, the graph model can be trained in the following manner: obtain the historical interaction records between any two of at least one advertiser, at least one candidate promoter, and at least one user; establish a graph network according to the historical interaction records, where the graph network includes multiple nodes, and the nodes are used to represent the advertiser, candidate promoter, or user. In the graph network, the connection edge connecting any two nodes is used to represent the interaction record between the two nodes in the historical interaction record, or is used to represent that through the historical interaction record, it is predicted that the two nodes can generate an interaction record in the future; based on the graph network, train the graph neural network model to obtain the graph model.

[0044] In this embodiment, the historical interaction records can refer to the above related embodiments, and this embodiment will not elaborate here.

[0045] Figure 3 is a schematic diagram of a graph network shown according to an embodiment of the present disclosure. Refer to Figure 3 , the nodes in the graph network represent merchants, users, or influencers.

[0046] In the graph network, the connection edge connecting any two nodes is used to represent the interaction record between the two nodes in the historical interaction record. For example, refer to Figure 3 , if there is a cooperation record between influencer 1 and merchant 1 in the historical interaction record, then the node corresponding to influencer 1 and the node corresponding to merchant 1 can be connected; for another example, continue to refer to Figure 3 , if there is a conversion record between user 1 and merchant 1 in the historical interaction record, that is, user 1 purchased the product of merchant 1, then the node corresponding to user 1 and the node corresponding to merchant 1 can be connected; for another example, continue to refer to Figure 3 , if in the historical interaction record, user 1 watched the content published by influencer 1 and user 1 purchased the product of merchant 1, then the node corresponding to user 1 and the node corresponding to influencer 1 can be connected; In the graph network, the connection edge connecting any two nodes can also be used to represent that through the historical interaction record, it is predicted that the two nodes can generate an interaction record in the future. For example, refer to Figure 3, in the case where user 1 browses the content published by influencer 1 and purchases the product corresponding to the content, assuming that influencers 2 and 3 also publish content about the product, it can be predicted that the content published by influencers 2 and 3 can also reach user 1, that is, it is predicted that user 1 can see the content published by influencers 2 and 3. Then, the node corresponding to user 1 can be connected to the node corresponding to influencer 2, and the node corresponding to user 1 can be connected to the node corresponding to influencer 3.

[0047] In this embodiment, the training of the graph model can refer to the following related embodiments, which will not be elaborated here.

[0048] In the above manner, the two nodes corresponding to the interaction record are directly connected through the interaction record recorded in the historical interaction record; in addition, reasonable prediction can be made based on the interaction record in the historical interaction record, and connection edges are constructed for the two nodes that are predicted to be able to generate interaction records in the future. In this way, the data volume of the graph network can be increased, and further the generalization of the graph model trained based on the graph network can be improved.

[0049] In some embodiments, the steps of training the graph neural network model based on the graph network to obtain the graph model may include the following steps: generating a node sequence based on the graph network; extracting positive samples and negative samples from the node sequence; and iterating the loss function of the graph neural network model based on the positive and negative samples to obtain the graph model.

[0050] In this embodiment, a node sequence can be generated based on random walk. Specifically, the graph network is randomly walked multiple times, starting from each node, randomly selecting its adjacent nodes for transfer, and repeating for several steps to obtain the node sequence. On this basis, for the target node in the node sequence, positive and negative samples corresponding to the target node are generated; and the loss function of the graph neural network model is iterated based on the positive and negative samples corresponding to the target node to obtain the graph model.

[0051] Among them, the length of the node sequence can be set according to the actual situation, which will not be elaborated here.

[0052] Among them, the positive and negative samples of the target node can be determined by the window size, and the window size represents the maximum number of nodes within the window. As an example, for the node sequence [A, B, C, D, E, F, G], taking the window size of 2 as an example, if the target node is node A, node B can be used as the positive sample of node A, and nodes C, D, E, F, and G can be used as the negative samples of node A. If the target node is node B, nodes A and C can be used as the positive samples of node B, and nodes D, E, F, and G can be used as the negative samples of node B.

[0053] Among them, the loss function includes the similarity loss of positive samples and the dissimilarity loss of negative samples. As an example, the loss function can be characterized by the following formula (2): (2) In the above formula (2), is the loss value, q is the number of positive sample nodes of the target node, p is the number of negative sample nodes of the target node, is the logarithmic function, is the Sigmoid function. The Sigmoid function is used to map the dot product to a probability value, is the vector representation of the target node, is the vector representation of the i-th positive sample node of the target node, is the vector representation of the c-th negative sample node of the target node.

[0054] Among them, the vector representation of the node can be updated through the backpropagation of the loss function, so that the vector representations of positive samples are closer to each other, and the vector representations of negative samples are farther from each other.

[0055] Among them, the stopping condition of the iteration can be that the number of iterations reaches a preset number, the descending gradient of the loss function is less than a preset gradient value, etc. This embodiment does not limit this.

[0056] In some embodiments, the connecting edges are given weights, and these weights can be used in the training of the graph model. In this case, the steps of training the graph neural network model based on the graph network to obtain the graph model may include the following steps: generating a node sequence based on the graph network and at least based on the weights. Among them, in the process of generating the node sequence, when determining the next node of the current node in the node sequence, the node with a larger weight among the nodes having a connecting edge with the current node has a greater probability of being determined as the next node; for the target node in the node sequence, generating positive and negative samples corresponding to the target node; iterating the loss function of the graph neural network model based on the positive and negative samples corresponding to the target node to obtain the graph model.

[0057] Among them, the weight can represent the association strength between nodes. In the method of using random walk to generate a node sequence, the weight of the edge can affect the walking process. For example, an edge with a larger weight is more likely to be selected by the random walk process, which means that the nodes connected by an edge with a larger weight are more likely to appear together in the node sequence. For example, continuing to refer to Figure 3 the graph network shown, if the weight of the connecting edge between influencer 1 and user 1 is 1, the weight of the connecting edge between influencer 2 and user 1 is 2, the weight of the connecting edge between influencer 3 and user 1 is 3, and the weight of the connecting edge between user 1 and merchant 1 is 4, when user 1 is the current node, the probability that merchant 1 is used as the next node of user 1 is the largest.

[0058] Among them, for determining positive and negative samples and the loss function, reference can be made to the above related embodiments, which will not be elaborated herein.

[0059] In the above manner, considering the influence of the weights of the edges between nodes on learning the vector representations of nodes, the weights are applied to the training of the model to improve the accuracy of the vector representations learned by the model.

[0060] In some embodiments, the weights of the connecting edges can also be used to adjust the contributions of positive and negative samples in the loss function. For example, for node pairs with larger weights of connecting edges, larger weights can be given to ensure that the vector representations of these nodes are closer during the model learning process, which can be achieved by introducing weights for characterizing the contribution degrees in the loss function.

[0061] In some embodiments, when generating positive and negative samples, the weights of the connecting edges can be used to adjust the distribution of the samples. For example, node pairs with larger weights of connecting edges can be preferentially selected as positive samples, while node pairs with smaller weights of connecting edges can be used as negative samples, which helps the model better learn the similarities and dissimilarities between nodes.

[0062] In some embodiments, when generating a node sequence, the variation law of the node types in the node sequence can also be controlled. For example, it is required that the types of two adjacent nodes in the node sequence are different. For another example, in consecutive nodes, the order of the node types is fixed. It should be noted that in the graph network, users, merchants, and influencers respectively correspond to three different types of nodes. Thus, the generation method of the node sequence can be determined according to actual needs.

[0063] In some embodiments, when the connecting edge between two nodes is used to represent the interaction record between the two nodes in the historical interaction record, the magnitude of the weight of the connecting edge connecting the two nodes is positively correlated with the number of interactions between the two nodes.

[0064] It should be noted that since the weights will be used in the training of the graph model, and the more the number of interactions, the stronger the correlation and relevance between the nodes are represented. Therefore, the magnitude of the weight of the connecting edge connecting the two nodes being positively correlated with the number of interactions between the two nodes helps the model better learn the similarities and dissimilarities between nodes.

[0065] In some embodiments, when the connecting edge between two nodes is used to represent predicting that an interaction record can be generated between the two nodes in the future through the historical interaction record, the magnitude of the weight of the connecting edge connecting the two nodes is determined by the prediction probability, and the prediction probability is used to characterize the probability that an interaction record can be generated between the two nodes in the future.

[0066] In this embodiment, the prediction probability can be the reach probability. Among two nodes with a connection edge, taking one node corresponding to an influencer and one node corresponding to a user as an example, the reach probability is used to represent the probability that the content promoted by the influencer can reach the corresponding user.

[0067] In some embodiments, the prediction probabilities corresponding to two nodes can be determined in the following manner: Obtain the interaction data of the two nodes within a preset time; according to the interaction data, use a pre-trained probability prediction model to determine the prediction probabilities corresponding to the two nodes.

[0068] Among them, the preset time can be selected according to the actual situation, and this embodiment does not limit this.

[0069] In this embodiment, among two nodes with a connection edge, taking one node corresponding to an influencer and one node corresponding to a user as an example, the interaction data can be the number of plays of the content promoted by the influencer on the user device corresponding to the user, the playing duration of the content promoted by the influencer on the user device corresponding to the user, the number of likes of the content promoted by the influencer by the user, the number of comments of the content promoted by the influencer by the user, whether the user has shared the content promoted by the influencer, whether the user has collected the content promoted by the influencer, and the number of times the user clicks on the homepage of the influencer through the user device, etc.

[0070] Among them, the pre-trained probability prediction model can be a decision tree, a random forest, xgboost, etc. The model can be trained based on historical interaction data, and then the probability prediction model can be obtained. On this basis, use the pre-trained probability prediction model to predict the prediction probabilities corresponding to the two nodes based on the interaction data.

[0071] Through the above method, the determination of the prediction probability is achieved.

[0072] In some embodiments, when constructing a graph network, a connection edge between two nodes can be constructed when the prediction probabilities of the two nodes are greater than or equal to a preset probability threshold, and the construction of a connection edge between the two nodes is prohibited when the prediction probabilities of the two nodes are less than the preset probability threshold. Among them, the preset probability threshold is selected according to the actual situation, and this embodiment does not limit this.

[0073] Since when the prediction probability is large enough, the correlation between nodes is higher, and when the model learns the vector representation based on the graph network, the distance between such nodes should be closer, while the distance between nodes corresponding to a smaller prediction probability should be farther. Therefore, a connection edge between two nodes is constructed only when the probability that the content promoted by the influencer can reach the corresponding user is large enough, so as to provide an accurate data basis for model training.

[0074] Figure 4 It is a block diagram of a device for determining a promoter shown according to an embodiment of the present disclosure. Refer toFigure 4 , the apparatus 400 for determining the promoter may include: An acquisition module 401, configured to acquire a first vector corresponding to a target advertiser and second vectors corresponding to at least one candidate promoter from a pre-trained graph model, where the graph model is generated based on historical interaction records between any two of at least one advertiser, the at least one candidate promoter, and at least one user, and is used to describe each advertiser, each candidate promoter, and each user through vectors, and the at least one advertiser includes the target advertiser; A first determination module 402, configured to determine the similarity between each of the second vectors and the first vector corresponding to the target advertiser; A second determination module 403, configured to determine a target promoter of the target advertiser from the multiple candidate promoters based on the similarity.

[0075] Optionally, the apparatus 400 further includes a training module, and the training module includes: A first acquisition sub-module, configured to acquire historical interaction records between any two of the at least one advertiser, the at least one candidate promoter, and the at least one user; A building sub-module, configured to build a graph network according to the historical interaction records, where the graph network includes multiple nodes, and the nodes are used to represent the advertiser, the candidate promoter, or the user. In the graph network, a connection edge connecting any two nodes is used to represent that there is an interaction record between the two nodes in the historical interaction records, or is used to represent that through the historical interaction records, it is predicted that the two nodes can generate an interaction record in the future; A training sub-module, configured to train a graph neural network model based on the graph network to obtain the graph model.

[0076] Optionally, the connection edge is given a weight, and the training sub-module is further configured to: Generate a node sequence based on the graph network, at least based on the weight. In the process of generating the node sequence, when determining the next node of the current node in the node sequence, the node with a larger weight among the nodes having a connection edge with the current node is more likely to be determined as the next node; Generate positive and negative samples corresponding to a target node in the node sequence; Iterate the loss function of the graph neural network model based on the positive and negative samples corresponding to the target node to obtain the graph model.

[0077] Optionally, when the connecting edge between the two nodes is used to represent the existence of an interaction record between the two nodes in the historical interaction record, the magnitude of the weight of the connecting edge connecting the two nodes is positively correlated with the number of interactions between the two nodes.

[0078] Optionally, when the connecting edge between the two nodes is used to represent that, based on the historical interaction record, it is predicted that the two nodes will be able to generate an interaction record in the future, the magnitude of the weight of the connecting edge connecting the two nodes is determined by a prediction probability, and the prediction probability is used to characterize the probability that the two nodes will be able to generate an interaction record in the future.

[0079] Optionally, the apparatus 400 further includes a third determination module, and the third determination module includes: a second acquisition sub-module, configured to acquire the interaction data between the two nodes within a preset time; a first determination sub-module, configured to determine the prediction probability corresponding to the two nodes according to the interaction data by using a pre-trained probability prediction model.

[0080] Optionally, the degree of attention received by the candidate promoter is greater than a preset degree of attention Among them, the implementation manners of the various modules of the above-mentioned apparatus 400 for determining the promoter may refer to the above-mentioned related embodiments, and are not described herein again in this embodiment.

[0081] An embodiment of the present disclosure further provides a computer-readable medium, on which a computer program is stored, and when the computer program is executed by a processing device, the steps of the above-mentioned method for determining the promoter are implemented.

[0082] An embodiment of the present disclosure further provides a computer program product, including a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned method for determining the promoter are implemented.

[0083] An embodiment of the present disclosure further provides an electronic device, including: a storage device, on which a computer program is stored; a processing device, configured to execute the computer program in the storage device to implement the steps of the above-mentioned method for determining the promoter.

[0084] Next, with reference to Figure 5 , which shows a schematic structural diagram of an electronic device 500 suitable for implementing an embodiment of the present disclosure. The terminal device in the embodiment of the present disclosure may include, but is not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Tablet Computers), PMPs (Portable Multimedia Players), in-vehicle terminals (such as in-vehicle navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 5The electronic device shown is merely an example and should not impose any limitation on the functions and usage scope of the embodiments of the present disclosure.

[0085] As Figure 5 shown, the electronic device 500 may include a processing device (such as a central processing unit, a graphics processing unit, etc.) 501, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 502 or the program loaded from the storage device 508 into the random access memory (RAM) 503. In the RAM 503, various programs and data required for the operation of the electronic device 500 are also stored. The processing device 501, the ROM 502, and the RAM 503 are connected to each other through a bus 504. The input / output (I / O) interface 505 is also connected to the bus 504.

[0086] Generally, the following devices may be connected to the I / O interface 505: an input device 506 including, for example, a touch screen, a touchpad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 507 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 508 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 509. The communication device 509 can allow the electronic device 500 to communicate with other devices wirelessly or wiredly to exchange data. Although Figure 5 the electronic device 500 with various devices is shown, it should be understood that it is not required to implement or have all the shown devices. More or fewer devices may be implemented or had alternatively.

[0087] Specifically, according to the embodiments of the present disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, the embodiments of the present disclosure include a computer program product, which includes a computer program carried on a non-transitory computer-readable medium, and the computer program contains program codes for executing the methods shown in the flowcharts. In such an embodiment, the computer program can be downloaded and installed from the network through the communication device 509, or installed from the storage device 508, or installed from the ROM 502. When the computer program is executed by the processing device 501, the above-mentioned functions defined in the methods of the embodiments of the present disclosure are executed.

[0088] It should be noted that the above-mentioned computer-readable medium in the present disclosure can be a computer-readable signal medium, a computer-readable storage medium, or any combination of the two. A computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples of computer-readable storage media can include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present disclosure, a computer-readable storage medium can be any tangible medium that contains or stores a program, which can be used by or in conjunction with an instruction execution system, apparatus, or device. In the present disclosure, a computer-readable signal medium can include a data signal propagated in a baseband or as part of a carrier wave, which carries computer-readable program code. Such a propagated data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on a computer-readable medium can be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0089] In some embodiments, the electronic device can communicate using any currently known or future-developed network protocol such as HTTP (HyperText Transfer Protocol), and can be interconnected with digital data communication in any form or medium (e.g., a communication network). Examples of communication networks include local area networks ("LAN"), wide area networks ("WAN"), the Internet (e.g., the Internet), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future-developed networks.

[0090] The above-mentioned computer-readable medium can be included in the above-mentioned electronic device; or it can exist separately without being assembled into the electronic device.

[0091] The above computer-readable medium carries one or more programs which, when executed by the electronic device, cause the electronic device to: obtain a first vector corresponding to a target advertiser and second vectors corresponding to at least one candidate advertiser respectively from a pre-trained graph model, wherein the graph model is generated based on historical interaction records between any two of at least one advertiser, the at least one candidate advertiser, and at least one user, and is used to describe each advertiser, each candidate advertiser, and each user through vectors, and the at least one advertiser includes the target advertiser; determine the similarity between each of the second vectors and the first vector corresponding to the target advertiser; and based on the similarity, determine a target advertiser of the target advertiser from the multiple candidate advertisers.

[0092] Computer program code for performing the operations of the present disclosure may be written in one or more programming languages or combinations thereof. The programming languages include, but are not limited to, object-oriented programming languages such as Java, Smalltalk, C++, and also include conventional procedural programming languages such as the "C" language or similar programming languages. The program code may execute entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., through the Internet using an Internet service provider).

[0093] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of systems, methods, and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagram may represent a module, a program segment, or a part of code that contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than marked in the accompanying drawings. For example, two consecutive blocks shown may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and the combinations of blocks in the block diagram and / or flowchart, may be implemented by a dedicated hardware-based system for performing the specified functions or operations, or may be implemented by a combination of dedicated hardware and computer instructions.

[0094] The modules involved in the embodiments of the present disclosure can be implemented in software or in hardware. In some cases, the names of the modules do not constitute a limitation on the modules themselves.

[0095] The functions described above herein can be performed, at least in part, by one or more hardware logic components. By way of example, and without limitation, the types of hardware logic components that may be used include: Field Programmable Gate Arrays (FPGA), Application Specific Integrated Circuits (ASIC), Application Specific Standard Products (ASSP), System on Chip (SOC), Complex Programmable Logic Devices (CPLD), and the like.

[0096] In the context of the present disclosure, a machine-readable medium may be a tangible medium that can contain or store a program for use by or in connection with an instruction execution system, apparatus, or device. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a Read-Only Memory (ROM), an Erasable Programmable Read-Only Memory (EPROM or Flash memory), an optical fiber, a portable Compact Disc Read-Only Memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0097] The above description is only a preferred embodiment of the present disclosure and an explanation of the applied technical principles. Those skilled in the art should understand that the scope of the disclosure involved in the present disclosure is not limited to the technical solutions formed by the specific combination of the above technical features, and should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above disclosure concept. For example, the technical solutions formed by mutually replacing the above features with the technical features (but not limited to) having similar functions disclosed in the present disclosure.

[0098] In addition, although the operations are depicted in a particular order, this should not be construed as requiring that the operations be performed in the particular order shown or in sequential order. In certain circumstances, multitasking and parallel processing may be advantageous. Similarly, although several specific implementation details are included in the above discussion, these should not be construed as limiting the scope of the present disclosure. Certain features described in the context of separate embodiments may also be implemented combinatorially in a single embodiment. Conversely, the various features described in the context of a single embodiment may also be implemented separately or in any suitable sub-combination in multiple embodiments.

[0099] Although the subject matter has been described in language specific to structural features and / or methodological logical acts, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or acts described above. On the contrary, the specific features and acts described above are merely example forms for implementing the claims. Regarding the apparatus in the above embodiments, the specific manner in which each module performs operations has been described in detail in the embodiments related to the method, and will not be elaborated here.

Claims

1. A method for determining a promoter, characterized in that: include: Obtaining a first vector corresponding to a target delivery party and a second vector corresponding to at least one candidate promoter from a pre-trained graph model, wherein the graph model is generated based on historical interaction records between at least one delivery party, at least one candidate promoter, and at least one user, and is used to describe each delivery party, each candidate promoter, and each user through a vector, wherein the at least one delivery party includes the target delivery party; Determine the similarity between each of the second vectors and the first vector corresponding to the target delivery party; Based on the similarity, a target promoter of the target delivery party is determined from the multiple candidate promoters.

2. The method according to claim 1, characterized in that The graph model is trained in the following way: Acquire historical interaction records between the at least one publisher, the at least one candidate promoter, and the at least one user; A graph network is established based on the historical interaction records, wherein the graph network includes a plurality of nodes, each node is used to represent the publisher, the candidate promoter or the user, and an edge connecting any two nodes in the graph network is used to indicate that there is an interaction record between the two nodes in the historical interaction record, or is used to indicate that, based on the historical interaction record, it is predicted that the two nodes will generate an interaction record in the future; Based on the graph network, a graph neural network model is trained to obtain the graph model.

3. The method according to claim 2, characterized in that The connecting edges are weighted, and the graph neural network model is trained based on the graph network to obtain the graph model, including: According to the graph network, a node sequence is generated based at least on the weight, wherein in the process of generating the node sequence, when determining the next node of a current node in the node sequence, a node with a greater weight among the nodes having the connection edge with the current node has a greater probability of being determined as the next node; For a target node in the node sequence, generating positive and negative samples corresponding to the target node; Based on the positive and negative samples corresponding to the target node, the loss function of the graph neural network model is iterated to obtain the graph model.

4. The method according to claim 3, characterized in that When the connection edge between the two nodes is used to indicate that there is an interaction record between the two nodes in the historical interaction record, the weight of the connection edge connecting the two nodes is positively correlated with the number of interactions between the two nodes.

5. The method according to claim 3, characterized in that: When the connecting edge between the two nodes is used to represent the prediction that the two nodes will be able to generate interaction records in the future through the historical interaction records, the size of the weight of the connecting edge connecting the two nodes is determined by the predicted probability, and the predicted probability is used to characterize the probability that the two nodes will be able to generate interaction records in the future.

6. The method according to claim 5, characterized in that The predicted probabilities corresponding to the two nodes are determined in the following way: Obtaining interaction data between the two nodes within a preset time; According to the interaction data, a pre-trained probability prediction model is used to determine the prediction probabilities corresponding to the two nodes.

7. The method according to any one of claims 1 to 6, characterized in that: The attention level of the candidate promoter is greater than a preset attention level.

8. A device for determining a promoter, characterized in that: include: an acquisition module, configured to acquire a first vector corresponding to a target delivery party and a second vector corresponding to at least one candidate promoter from a pre-trained graph model, wherein the graph model is generated based on historical interaction records between at least one delivery party, at least one candidate promoter, and at least one user, and is configured to describe each delivery party, each candidate promoter, and each user through a vector, wherein the at least one delivery party includes the target delivery party; A first determination module, used to determine the similarity between each of the second vectors and the first vector corresponding to the target delivery party; The second determining module is used to determine a target promoter of the target delivery party from the multiple candidate promoters based on the similarity.

9. A computer readable medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processing device, the steps of the method according to any one of claims 1 to 7 are implemented.

10. An electronic device, characterized in that: include: a storage device having a computer program stored thereon; A processing device, configured to execute the computer program in the storage device to implement the steps of the method according to any one of claims 1 to 7.

11. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images